摘要： “When children read more, they learn to read better,” says Dominic Wyse at University College London. A young child who enjoys being read to is developing a love of books that will help motivate them to learn to read.

Researchers are finally getting to grips with how children's brains pick up reading. Now the challenge is to apply this to the classroom to help kids reach their potential in literacy

MORE than 5000 years after the invention of writing, you would think we would have completely nailed the best way to teach people to read – literacy is a key skill in most societies, after all. But you would be wrong. Not only have scientists long disagreed on the most effective methods, but their arguments have fuelled a decades-long, politicised “reading war” over how to teach children to read English.

Meanwhile, large numbers of children are struggling to achieve the standards they should for their age. Last year, just 33 per cent of 9 and 10-year-olds in the US were assessed as being proficient or advanced readers. “The US has done poorly in teaching kids to read for a long time,” says Mark Seidenberg, a cognitive neuroscientist at the University of Wisconsin-Madison. And the problem isn’t confined to English-speaking countries: there is also confusion about how to teach children to read other languages.

A key battleground is a teaching technique known as phonics. In the US, poor literacy is often blamed on having too little phonics in the classroom. But, confusingly, researchers last year argued that children in England are being failed by being taught too much phonics. Herein lies the root of the problem: it is one thing to understand how kids learn to read and, it turns out, quite another to figure out how best to teach them. The good news is that researchers, having begun to enter the classroom, are finally getting to grips with how they can translate their insights to improve teaching and, ultimately, bring an end to the reading wars.

For 400 years or more, there has been debate about the best way to teach reading. Today, most reading researchers say the learning process begins when the student – generally assumed to be a child – realises that the words they speak can be represented by text. A child learning an alphabetic script like English must also recognise that spoken words can be broken down into small elements called phonemes – the /k/ sound in “cat”, for instance – and that the alphabetic letters, or graphemes, represent these phonemes. With time, the child learns to associate each grapheme with one or more phonemes, and to blend strings of phonemes together to make words. They can then turn a written word, such as cat, first into a set of discrete phonemes – /k/a/t/ – and then into the spoken word “cat”. Researchers call this decoding. Formal instruction in decoding, otherwise known as phonics, allows children to read many written words. Ultimately, those words become familiar enough that the child recognises them instantly and knows how to pronounce them without decoding.

Read more: More education is what makes people live longer, not more money

“Phonics has been a part of reading education in England since the 19th century,” says Dominic Wyse, who studies early education at University College London (UCL). But in 2006, after the publication of a government-commissioned independent review of teaching methods, phonics instruction gained a more dominant role. This is what is undermining reading education in England, Wyse and Alice Bradbury, also at UCL, concluded last year after surveying more than 2000 teachers.

The trouble is that there is much more to reading than just decoding. Take a simple written word like bow, which has several meanings and two distinct pronunciations. Decoding alone won’t help someone select the right combination of these. They must also use contextual cues from the text, such as the context contained in an entire sentence: “bow if you meet the king” versus “I can tie a bow”, for example. In other words, children need to understand what they read.

We have known for decades that reading instruction is less effective if children are taught to decode lists of words shorn of context. Better results are seen when they learn using books. In addition, it helps for teachers to emphasise that decoding is a tool to aid reading comprehension. Some reading researchers say teachers in England do this. But Wyse and Bradbury found that only about a quarter of the teachers in their study agreed that they did. Wyse thinks there are several reasons for the excessive focus on phonics, not least that young children are required to take a national test that assesses their ability to decode. “Teachers will teach to the test because there are implications if the children don’t pass,” he says.

The situation is different in the US. For years, teachers in many states have placed books at the heart of the reading curriculum and encouraged children to derive meaning from them. In recent years, however, prominent commentators have argued in news articles and podcasts that this approach is largely to blame for the reading crisis in the US. How can that be, given what the science says?

Balancing act

The problem is one of emphasis. The reading approach long favoured in many US schools is known as the “three-cueing system” because it instructs children to use three different strategies while reading. Two of those “cues” are semantics (meaning) and syntax (sentence structure): in essence, using contextual cues to check whether the words make sense. That fits with the science. However, children may then be encouraged to use this information to predict – or, more bluntly, to guess – what word comes next in a sentence. (The boy is flying his… what? Perhaps his kite?) Then comes the third cue: they look at the letters in that word and check whether or not they match their prediction.

This may sound curiously back-to-front. In fact, very few reading researchers recommend the three-cueing system because it puts decoding at the end of the reading process rather than at the beginning. However, some are equally concerned that the backlash against it has seen commentators suggest children shouldn’t be taught to use cues at all because they can get in the way of their reading development. This could lead to reading curricula in the US that contain phonics and little else, says Donna Scanlon at the University at Albany, New York. She points out that many schools experimented with such an approach in the early 2000s, during a federal education programme called Reading First. A government assessment of the programme was published in 2008. “Kids got better at phonics, but they didn’t get better at reading comprehension,” says Scanlon. Other researchers argue that methodological problems with the assessment make it far from clear that Reading First turned children into great decoders but poor readers. Nevertheless, Scanlon’s concerns do mirror those of Wyse and Bradbury.

SPRINGFIELD, VA - NOVEMBER 3: First grade teacher Emily VanDerhoff, who is a member of Fairfax County NAACPs education committee, teaches phonics-based reading lessons at Hunt Valley Elementary School on November 3, 2022, in Springfield, VA. Fairfax County NAACPs education committee has pushed Fairfax County Public Schools to re-introducing phonics-based reading instructions and interventions. (Photo by Jahi Chikwendiu/The Washington Post via Getty Images)

When learning to read, children from different backgrounds may have different needs

What is needed, say these researchers, is a balanced approach to reading instruction that encourages children to complement decoding with contextual cues while remembering that the goal of reading is to understand texts. Over the past 25 years, Scanlon and her colleagues have been working on exactly such a method, and their research suggests that the strategy is effective. However, it is specifically designed to help children identified as having difficulties with reading, in small groups or even one-to-one. We still don’t really know the best way to teach such skills to a classroom of children, for the simple reason that understanding the teaching of reading, as opposed to understanding how children learn to read, hasn’t always been a research priority.

“It’s one thing to say: we know how the brain learns how to read,” says Emily Solari at the University of Virginia. “It’s another thing to say: we know how to take that science and translate it into effective practices that teachers can actually implement.” What researchers like Solari are discovering is that getting an evidence-based answer to that question is far from easy. “It’s really hard to do that kind of research,” says Seidenberg. “There are so many factors floating around.”

Read more: How to hack your unconscious to boost your memory and learn better

For instance, it is widely accepted that hunger influences children’s ability to learn – which means that, in theory, a child may respond to the same lesson differently on a full or empty stomach. “Hunger is a huge component to engagement,” says Jessica Jang, who teaches first grade (6 and 7-year-olds) in a school in California. “When a student is hungry, this can result in misbehaviour, such as withdrawing from the lesson or causing disruptions.”

Teachers also need to recognise that children from different backgrounds may have different needs. Bryant Jensen at Brigham Young University, Utah, points to some research showing that Navajo Native American children tend to feel embarrassed if they are singled out for praise during reading lessons, which can undermine their willingness to participate in classroom activities.

What’s more, Patrick Proctor at Boston College in Massachusetts and Chris Chang-Bacon at the University of Virginia have noted that the standards used to assess reading in the US are usually based on the performance of monolingual English speakers. This means that a teacher may, for example, have a narrow sense of the correct pronunciation of a word, and assume a child who pronounces it differently is struggling with decoding.

Back to the classroom

Proctor says researchers must spend time in the classroom. By observing teachers and children and talking to them, they can go beyond asking whether a curriculum is effective and begin to understand why. Since 2014, he and his colleagues have been working in US schools to develop a literacy curriculum for multilingual students that plays to their strengths. For instance, it uses video to communicate information through images, action and sound to help them expand their vocabulary both in English and in the other language or languages they use – a vital skill for readers.

At a broader level, there is more awareness today that children face different demands depending on the language they are learning to read. For example, a strong focus on phonics makes sense when teaching children to read English, given the irregular spelling of many English words. But other languages, including German and Spanish, are far easier to decode so children don’t need to devote months or years to phonics lessons, says David Share at the University of Haifa in Israel. Then there are non-alphabetic languages where individual symbols may represent whole words rather than phonemes, which makes phonics instruction far less important.

Share believes that the reading crises in some parts of the world may reflect the progress made in understanding reading in English. The problem is that English-specific findings might influence reading education policy in countries using writing systems for which they aren’t applicable. Sonali Nag at the University of Oxford has come across many examples of this. One is the assumption that children will learn all of the characters of a writing system in their first school year – which may be true of an alphabetic writing system with a few dozen symbols, but not for non-alphabetic writing systems with hundreds or even thousands of symbols. “It’s blindingly obvious, but it’s not in the theoretical models,” she says.

For Nag, the solution is to ramp up research into understudied writing systems so that educational policies can be based on relevant reading science. Her work focuses on the “alphasyllabic” writing systems of South and South-East Asia. Unlike the alphabets common to Europe and North America, here the units used to construct words – called akshara – are usually based on a single consonant that is modified by a smaller vowel symbol. This means akshara often represent syllables rather than phonemes. Nag has begun examining children’s literature across Asia – such as that written in the Kannada script used in south India – to better understand the range of words encountered by young readers in this region.

Read more: Dyslexia linked to 42 genetic variants in biggest study of its kind

Taking this international perspective is a win-win. In a recent paper, Share explained how studies in non-English reading can also advance English reading research. For instance, researchers studying reading acquisition in easy-to-decode languages, such as German, tend to use different measures – particularly fluency – to assess how well children are learning. Partly as a result of this, fluency has become widely recognised as an important measure of reading performance in English. It has even influenced the way scientists think about dyslexia, which they now see as including problems with fluency and not just problems with decoding words.

Despite talk of reading crises in the US and elsewhere, in recent decades, reading research has greatly increased our understanding of how children learn this key skill. That is good news for parents and caregivers who want to help their children reach their full potential (see “Raising a reader”). Clearly, though, there is still plenty to do – even if it appears that reading researchers now seem to be slowly moving in the right direction. “Translation of evidence into practice takes a really long time,” says Solari. “If this were easy, we would already have figured it out.”